Content
61%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The reference material, examples, and tips are genuinely useful and specific, but the opening Purpose/Usage sections are redundant with the description, the label-fetching step lacks an executable command, and the workflow has no validation checkpoint. Overall a solid, actionable skill that would benefit from tightening and one verification step.
Suggestions
Collapse the redundant "Purpose" and "Usage" sections into one or two sentences and drop marketing filler like "leverage GitHub's powerful search capabilities".
Make step 2 executable by naming the command, e.g. `gh api repos/{owner}/{repo}/labels --jq '.[].name'` or `gh label list --repo owner/name`.
Add a final validation step, e.g. verify every term in the constructed query appears in the Search Syntax Overview and that chosen labels actually exist in the fetched label list before presenting the query.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The core material (syntax overview, examples, tips) is dense and earns its tokens, but the "Purpose" and "Usage" sections largely restate each other and the frontmatter description ("This skill takes a natural language query... and converts it into a properly formatted GitHub search query" vs. "provide a natural language query... the skill will then analyze the input"), plus filler like "leverage GitHub's powerful search capabilities". Not 2: padding is limited to the two intro sections, not pervasive. | 3 / 5 |
Actionability | The tips are concrete and directive ("Always include a 'sort:' parameter", "Use a '-' in front of a syntax component to indicate that it should be 'not-ed'") and the seven example queries are copy-paste ready. Not 5: step 2 ("Fetch labels for the repo if mentioned") gives no executable command (e.g. a `gh api` or `gh label list` invocation), leaving a gap in how to execute the workflow. | 4 / 5 |
Workflow Clarity | The three-step sequence (identify repo → fetch labels → convert per tips) is present but has no checkpoints or validation — nothing instructs verifying the formed query against the syntax overview or confirming label matches before presenting it. This is not a destructive/batch operation so no hard cap applies, but it fits "steps listed but validation gaps" better than "most checkpoints present". | 3 / 5 |
Progressive Disclosure | The body is a single, well-organized file with clear sections (Steps, Search Syntax Overview, Example Queries, Tips, Adding Labels) and no nested references; no bundle files exist and none are strictly needed at this size. Not 5: the ~34-line syntax key-value table is reference material that could live in a references/ file, keeping SKILL.md as a lean overview; not 3: the structure is clean and the content is not misplaced at scale. | 4 / 5 |
Total | 14 / 20 Passed |